ABSTRACT

With the widespread use of tag clouds, multiple map-based variations have been proposed. Like standard tag clouds (also called word clouds), these ‘tag maps’ all share the basic strategy of displaying words within a ‘geographic space’ and scaling the word size to depict frequency (or importance) of those words within some dataset. While some tag maps simply plot a standard tag cloud on top of a map, the subset of tag maps we focus on here are those in which the collection of words are displayed within bounded geographic regions (often of irregular shape) that the words are relevant for. For this form of tag map, map scale and polygon shape add constraints to word size and position that have not been considered in most prior approaches to tag map word layout. In this paper, we present a layout strategy for tag map generation that includes consideration of the shape and size of the geographical regions acting as containers for the tags. The method introduced here uses a triangulated irregular network (TIN) to subdivide the geographical region into many triangle subareas, with the centroid of each triangle being a potential location to centre a tag on. All the triangles are sorted by their area and all the tags are sorted by their weight value (e.g. frequency, importance or popularity). Positioning of tags is undertaken sequentially from most important (or frequent or popular) with potential locations being the TIN triangle centroids (tried from largest to smallest triangle). After each tag placement, the TIN is recalculated to integrate the tag centroid and bounding corners into the TIN creation. The limited whitespace in the geographical region, at any specific scale, is used fully by dynamically adjusting the font size along with the number and the direction of tags. The method can be applied to add tags within geographic polygons that are convex, concave and other more complex regions containing holes or islands.

摘要

随着标签云的广泛使用，已经提出了多种基于地图的变体。像标准标签云（也称为词云）一样，这些“标签图”都具有基本的策略，即在“地理空间”内显示单词并缩放单词大小以描述某些数据集中这些单词的频率（或重要性）。尽管有些标签图只是在地图顶部绘制了一个标准标签云，但我们这里重点关注的标签图子集是其中单词集合显示在与单词相关的有限地理区域（通常为不规则形状）中对于。对于这种形式的标签地图，地图比例尺和多边形形状增加了字大小和位置的约束，而在大多数现有的标签地图字布局方法中都没有考虑这些约束。在本文中，我们提出了一种用于标签地图生成的布局策略，其中包括考虑充当标签容器的地理区域的形状和大小。此处介绍的方法使用三角不规则网络（TIN）将地理区域细分为许多三角形子区域，每个三角形的质心是将标签居中的潜在位置。所有三角形均按其面积排序，所有标签均按其权重值（例如频率，重要性或受欢迎程度）排序。标签的定位是从最重要（或频繁或流行）开始顺序进行的，潜在位置是TIN三角形形心（从最大到最小的三角形尝试）。每次放置标签后，将重新计算TIN，以将标签的质心和边界角整合到TIN创建中。通过动态调整字体大小以及标签的数量和方向，可以充分利用地理区域中任何特定比例的有限空白。该方法可以应用于在凸，凹和包含孔或岛的其他更复杂区域的地理多边形内添加标签。